the Galvanic Current through Iron. 5 



whatever kind they may be) elimination furnishes very reli- 

 able results. Thereby the scheme of all the experiments which 

 I made assumed the form a 1 — b — a 2 , where a 1 signifies the 

 experiment before the magnetization, b the same after magne- 

 tizing, and a 2 after demagnetizing. The results b and 1 2 



a 



are then fairly comparable if the difference a l — a 2 is small. 



A second portion of the temperature-influences, that occa- 

 sioned by the surroundings, can also be reduced to a minimum, 

 and that minimum eliminated. The latter can here be done 

 with peculiar facility, since the periods of the extreme varia- 

 tions of temperature stand in no connexion whatever with 

 those of the alterations of the resistance conditioned by the 

 experiments. 



There still remains for discussion the third and most import- 

 ant part of the influences of temperature, namely those condi- 

 tioned by the magnetizing-arrangement. This in the beginning 

 consisted of a spiral of thick wire, spun over and waxed, and 

 wound upon a glass tube, through which the magnetizing 

 current flowed. The iron wire was pushed into the glass tube, 

 and the entire apparatus set up at a distance of several metres 

 from the galvanometer ; so that neither the magnetism nor 

 even the magnetizing current acted directly upon the galva- 

 nometer. It nevertheless appeared that the heating action of 

 the magnetizing current was conveyed by radiation from the 

 spiral to the iron wire and occasioned variations in its resist- 

 ance, which, partly by their considerable amount, and partly 

 because the periods of both variations are identical, concealed 

 all the alterations which were to be measured*. Hence it 

 was important to construct an adiathermanous magnetizing- 

 apparatus. In this I succeeded by making use of the proce- 

 dure often employed by Joule in his thermal investigations — 

 namely, placing in layers one over another a number of series, 

 each consisting of an adiathermanous, a badly conducting 

 material, and a material of great heat-capacity. The copper 

 spiral was accordingly wound upon a wide glass tube, this 

 drawn over a thick-walled caoutchouc hose ; inside this a nar- 

 rower glass tube was inserted, and in the latter the iron wire, 

 mostly wrapped round with paper. To carry out the principle 

 completely I should have had to put between the narrower 

 glass tube and the iron wire a layer of substance of great heat- 

 capacity — for example, a stationary stream of alcohol or melt- 



* These variations of resistance furnish a very serviceable method of 

 following the temperature in a galvanic battery. I intend to return to 

 this subject in another place. 



